<p>VoicePro: An Artificial Intelligence Voice Assistant for Integrated Personal Productivity Management</p>

Many knowledge workers and students often find it challenging to manage personal productivity across digital tools. Most productivity applications available today only cater to one aspect of workflow management: task tracking, time management, or analytics, which causes context switching. Moreover, most productivity software has a manual one-to-one GUI approach that does not leverage the power of voice-based natural language interfaces to be more efficient. VoicePro is a comprehensive AI-powered productivity tool that is built on the advanced capabilities of GPT-4o-mini. It merges task management, Pomodoro-style focus timers, calendar scheduling, reminders, and productivity analytics in one single web app with the help of a voice assistant. VoicePro, powered by the advanced Whisper STT model of OpenAI, processes speech-to-text transcription, and the advanced GPT-4o-mini model provides natural language intent recognition. Users can perform a range of productivity operations through simple conversational voice commands. The system adopts a hybrid voice processing framework using a deterministic fallback parser to ensure reliability in case of unavailability of AI services. The design is a three-tier client-server architecture employing a React 18 single-page application as the front end, a REST API by Flask with JWT authentication, and Google OAuth 2.0 as the back end. We study VoicePro in three ways. We compare the time taken to complete tasks with voice and manual input modalities. Then, we measure the accuracy of command recognition from test utterances. According to the preliminary results, we found that users take 42% less time to create tasks through voice versus filling forms. Moreover, the percentage accuracy of the AI intent parser was 88% across six command categories.